Group 2 Innate Lymphoid Cells Are Detrimental to the Control of Infection with Francisella tularensis

Key Points Enhancing ILC2 numbers increases Francisella tularensis bacterial burden. Reducing ILC2 numbers promotes Francisella tularensis bacterial clearance. IFN-γ reduces ILC2s, with ILC2 production of IL-5 detrimental to bacterial clearance.

Innate lymphoid cells (ILCs) are capable of rapid response to a wide variety of immune challenges, including various respiratory pathogens. Despite this, their role in the immune response against the lethal intracellular bacterium Francisella tularensis is not yet known. In this study, we demonstrate that infection of the airways with F. tularensis results in a significant reduction in lung type 2 ILCs (ILC2s) in mice. Conversely, the expansion of ILC2s via treatment with the cytokine IL-33, or by adoptive transfer of ILC2s, resulted in significantly enhanced bacterial burdens in the lung, liver, and spleen, suggesting that ILC2s may favor severe infection. Indeed, specific reduction of ILC2s in a transgenic mouse model results in a reduction in lung bacterial burden. Using an in vitro culture system, we show that IFN-g from the live vaccine straininfected lung reduces ILC2 numbers, suggesting that this cytokine in the lung environment is mechanistically important in reducing ILC2 numbers during infection. Finally, we show Ab-mediated blockade of IL-5, of which ILC2s are a major innate source, reduces bacterial burden postinfection, suggesting that IL-5 production by ILC2s may play a role in limiting protective immunity. Thus, overall, we highlight a negative role for ILC2s in the control of infection with F. tularensis. Our work therefore highlights the role of ILC2s in determining the severity of potentially fatal airway infections and raises the possibility of interventions targeting innate immunity during infection with F. tularensis to benefit the host. The Journal of Immunology, 2023, 210: 618627.
F rancisella tularensis is a highly infectious intracellular bacterium and the causative agent of tularemia (1). Capable of extensive replication and dissemination to peripheral organs (2), respiratory infection with F. tularensis can cause lethal disease in the absence of intervention (3). Moreover, the enhanced virulence of the bacterium via this route has led to concerns that F. tularensis could be used as a biological weapon if aerosolized (4,5).
Much of our current knowledge of the early immune response against F. tularensis has focused on the role of myeloid cells (6). For example, macrophages and neutrophils serve as sites of extensive replication during respiratory infection, with dysregulation of these mechanisms proving detrimental to the host immune response (79). In contrast, the contributions of innate lymphocyte responses in the pathogenesis or control of infection with F. tularensis are less understood. Innate lymphoid cells (ILCs) are tissue-resident innate lymphocytes, which are capable of a rapid response to a wide variety of immunological challenges. ILCs are divided into three major subsets that is, ILC1s, ILC2s, and ILC3s, which are largely analogous to CD4 1 Th cell subsets in their phenotype and effector functions (10).
ILCs are frequently found at barrier sites, such as the skin, intestine, and the lung and play critical roles in early protection against invading pathogens (10). At steady state, ILC2s represent the dominant ILC subset in the murine lung (11), playing key roles in the immune response against several respiratory pathogens (1214). However, the role of ILCs during respiratory infection with F. tularensis is currently unknown. Thus, to address this gap, we used a murine model of infection with a live vaccine strain (LVS) of F. tularensis. This strain has been instrumental in the study of tularemia, as it represents a less virulent strain in humans when compared with other strains such as Schu S4 (1), while retaining its lethality in mice when administered via the intranasal (i.n.) route (15).
In this study, we demonstrate that F. tularensis LVS infection is associated with a loss of ILC2s from the lungs. Moreover, we reveal that this reduction in ILC2s may benefit the host immune response against the bacteria, as, conversely, expansion of ILC2 numbers by either IL-33 treatment or adoptive transfer of ILC2s causes significantly exacerbated bacterial burdens in the lung, liver, and spleen of mice. In further support of a reduction in ILC2s being beneficial for host immunity, depletion of ILC2s using a transgenic mouse model resulted in reduced bacterial burden in the lung. Mechanistically, we show that IFN-g from the LVS-infected lung appears to drive the infection-induced reduction in ILC2s. Finally, blockade of IL-5, of which ILC2s are a major innate source, reduces bacterial burden, suggesting that ILC2-produced IL-5 may limit protective host immunity. Overall, our data highlight a negative role for ILC2s in the control of infection with F. tularensis LVS. We therefore postulate that manipulation of ILC2 numbers and cytokines during early infection with F. tularensis could be of potential therapeutic benefit.

Materials and Methods
Bacterial strains F. tularensis LVS was derived from an original NDBR101 Pasteurella tularensis live vaccine, experimental lot 4. The vaccine contained 6 × 10 9 CFU of lyophilized F. tularensis LVS, which was stored in a culture collection at −20 • C at the Defense Science and Technology Laboratory. Master stocks and i.n. challenge doses were prepared as previously described (9).

Mice and in vivo F. tularensis infection
Female C57BL/6 mice (Charles River, Margate, U.K.), Rag1 −/− mice on a C57BL/6 background (16), inducible ICOSdiphtheria toxin receptor (DTR) (iCOS-T) mice on a C57BL/6 background (17), originally a gift from Dr. Andrew McKenzie (MRC Laboratory of Molecular Biology, Cambridge, U.K.), and Red-5 mice (Red5Cre (B6(C)-Il5 tm1.1(icre)Lky /J, stock number 030926), purchased from The Jackson Laboratory and originally generated by Prof. Richard Locksley, University of California San Francisco [18]) were kept in specific pathogen-free conditions according to institutional and U.K. Home Office guidelines in the Biological Services Unit at The University of Manchester. Eight-to 12-wk-old mice were infected i.n. with 50 ml of PBS containing ∼1000 CFU of F. tularensis LVS. An infectious dose was confirmed by CFU count on blood cysteine glucose agar plates. All procedures were performed in accordance with the Home Office Scientific Procedures Act (1986) and under the Department for Environment, Food and Rural Affairs license. Survival of mice was defined as mice reaching the defined humane endpoint as specified by our license granted by the U.K. Home Office. For all experiments, the humane endpoint was defined as either weight loss in excess of 25% of starting body weight, and/or loss of mobility and condition, as determined by use of a clinical scoring system that monitored the general condition of mice throughout experiments (9).

Depletion of ILC2s in iCOS-T mice
iCOS-T mice, which allow specific deletion of ILC2s via injection of DT (Sigma-Aldrich) (17), were treated via i.p. injection daily with 1 mg of DT/mouse. Mice were treated daily from 4 d prior to i.n. infection with F. tularensis LVS, and daily postinfection (p.i.) until mice were culled for analysis.
Treatment of mice with antiIL-5 Ab C57BL/6 mice were treated i.p. with 100 mg/mouse antiIL-5 Ab (clone TRFK5) or rat IgG1 isotype control (anti-HRP) (both from Bio X Cell). Mice were treated every other day from 4 d prior to infection prior to i.n. infection with F. tularensis LVS, and every other day p.i. until mice were culled for analysis.

Enumeration of bacterial burden
Bacterial burdens were enumerated from the lung, liver, and spleen of mice, with samples processed <2 h postmortem. All organs were collected in PBS and weighed. Samples were disrupted through a 40-mm cell sieve, followed by serial dilution of organ homogenates in PBS. Although the use of PBS in this step will not cause extensive cell lysis, this method was used to ensure that cells were available for downstream flow cytometry analysis and gave a consistently robust measure of bacterial CFU in organs. Bacteria were incubated for 45 d at 37 • C, and once grown sufficiently, single colonies were counted and organ weights were used to express data as CFU per gram of organ.

Digestion of lung tissue
Lung tissue was cut into small sections and shaken in an incubator at 37 • C in PBS containing 3 mg/ml collagenase D (Roche) for 40 min. After digest, tissues were disrupted through a 40-mm cell sieve and samples washed with PBS. RBCs were lysed using RBC lysing buffer Hybri-Max (Merck), washed again in PBS, and cell numbers were counted before staining and analysis by flow cytometry.
Gating was established using either fluorescence minus one or isotype controls. Prior to data acquisition, cells were washed in FACS buffer and stored at 4 • C until analysis on a BD LSRFortessa (BD Biosciences). Data were analyzed using FlowJo software version 10.7.1 (Tree Star, Ashland, OR).

Administration of IL-33
Recombinant murine IL-33 (rIL-33; carrier-free, R&D Systems) was administered to mice via the i.p. route at a dose of 0.5 mg/mouse every 2 d starting 5 d before infection with F. tularensis LVS. Control mice were administered vehicle (PBS).

In vitro ILC2 culture
Lung ILC2s were sort-purified as above, and cells were transferred to 96-well plates in 200 ml of complete RPMI 1640 at 1 × 10 5 ILC2s per well. Cells were then cultured for 7 d with IL-7 (Invitrogen) at 25 ng/ml. Naive and LVS-infected lung supernatants were generated by 24-h incubation of homogenized lung samples at 37 • C in complete RPMI 1640. Cell-free supernatants were collected by centrifugation and filtration through a sterile syringe 0.22-mm microfilter. Then, 100 ml of supernatants was then added to 1 × 10 5 ILC2s (in 100 ml) and cultured for 5 d. All wells were cultured with IL-7 at 25 ng/ml and, where indicated, antiIFN-g Ab (clone XMG1.2, Bio X Cell) was added to culture media at 10 mg/ml. For live ILC2 cell counts, a hemocytometer counting chamber was used, with live ILC2 cell counts enumerated via trypan blue dye exclusion.

Statistical analysis
All graphs and statistical analyses were produced using GraphPad Prism 9. Normality of data was determined using the ShapiroWilk normality test, with appropriate parametric and nonparametric statistical tests performed for each dataset (stated in figure legends). Data were expressed as mean ± SEM. Statistical significance was considered at p < 0.05.

ILC2s are reduced during infection with F. tularensis LVS
To define the ILC response during pulmonary infection with F. tularensis LVS, ILCs were identified in the lung and defined by a lack of expression of lineage markers (Lin − ) and high expression of CD90.2 and CD127 ( Fig. 1A; full gating strategy is shown in Supplemental Fig. 1). Interestingly, we observed a significant and continual decrease in total lung ILCs during the course of infection, both in frequency and absolute cell numbers (Fig. 1B, 1C).
To determine whether this reduction in total ILC numbers was a result in changes to a specific ILC subset, we further analyzed specific subgroups of ILCs. As previously reported (12), we found that ILC2s (defined as ST2 1 ) represented the dominant lung ILC subset under steady-state conditions, with a small number of ILC1s (T-bet 1 ) present (Fig. 1D). In contrast, ILC3s (RORgt 1 ) were virtually absent from the airways in both naive and infected mice (Fig. 1D). Upon infection with F. tularensis LVS, both the frequency and total number of ILC2s were significantly reduced from day 5 p.i. (Fig. 1DF). This subset-specific reduction was also associated with a smaller but reproducible concurrent increase in ILC1 numbers (Fig. 1D, 1G, 1H). Overall, these data indicate that pulmonary infection with F. tularensis LVS results in significant perturbation and reduction in the lung ILC compartment during the later stages of infection, predominantly due to a reduction in ILC2s.

IL-33induced expansion of the lung ILC2 compartment significantly enhances bacterial burden
ILC2s have previously been shown to be involved in the immune response against other intracellular infections (12, 13) and tissue repair p.i. (12). Thus, it was hypothesized that the subset-specific reduction in ILC2s could be advantageous to the growth and pathogenesis of F. tularensis LVS. To address how changes in ILC2s impact the progress of F. tularensis infection, mice were treated with IL-33 prior to i.n. infection (see Fig. 2A for treatment regimen). As expected (10), IL-33 treatment caused a significant expansion of ILC2s, both by percentage and total numbers, in the naive mouse lung (Fig. 2BD). This expansion was also observed in the infected lung, albeit to a lesser extent than in IL-33treated uninfected mice (Fig. 2BD), suggesting that infection antagonized the ILC2 response even following IL-33induced expansion.
We next tested how IL-33 treatment affected pathogen burdens and infectious outcome. Surprisingly, IL-33 treatment resulted in enhanced morbidity in mice infected with F. tularensis LVS, which required euthanasia of animals by day 4 p.i. under animal license restrictions (Fig. 2E). Moreover, treatment with IL-33 resulted in significantly enhanced bacterial burdens in the lung, liver, and spleen at day 4 p.i. (Fig. 2F). Taken together, these data indicate that IL-33 treatment is detrimental to the outcome of infection with F. tularensis LVS.
Our findings showed that the IL-33 treatment could exacerbate infection in the host, potentially via ILC2 expansion. However, as IL-33 is a pleiotropic cytokine that can affect other immune cells (18,19), it was important to consider the wider impact of IL-33 treatment on the immune system during F. tularensis infection. In line with ILC2 expansion and prior reports (18), we found that IL-33 treatment caused enhancement of eosinophil numbers (Supplemental Fig. 2AC). Furthermore, type 2 CD4 1 ST2 1 T cell (Supplemental Fig. 2D, 2E) and CD4 1 Foxp3 1 regulatory T cell numbers were also enhanced during infection (Supplemental Fig. 2F, 2G). Thus, to address whether the effects of IL-33 treatment were acting via enhancement of innate or adaptive immune responses, we used Rag1 −/− mice that lack adaptive immune cells. We observed that, even in the absence of adaptive immunity, IL-33 induced a significant increase in bacterial burdens in Rag1 −/− mice in the lung, liver, and spleen to the same extent to that seen in wild-type (WT) mice (Fig. 2G), with both ILC2s and eosinophils still significantly expanded in IL-33treated LVS-infected Rag1 −/− mice (Supplemental Fig. 3A, 3B). Thus, these data strongly indicated that the innate immune response was sufficient to promote IL-33mediated enhancement of bacterial burdens during F. tularensis LVS infection.
Next, to directly test the role for enhanced lung ILC2s to the outcome of infection, sort-purified ILC2s were transferred to WT mice via the i.n. route at day 3 p.i. (Fig. 3A). Strikingly, transfer of ILC2s also resulted in significantly enhanced bacterial burdens in the lung, liver, and spleen (Fig. 3B), similar in magnitude to that observed when mice were treated with IL-33 (Fig. 2F). Thus, together, these data strongly suggest that ILC2s are detrimental to the control of infection with F. tularensis LVS.

Depletion of ILC2 numbers significantly reduces pulmonary bacterial burdens after F. tularensis LVS infection
The data above show that promoting ILC2 numbers, either via injection of IL-33 or direct transfer of ILC2s to the lung, causes enhanced bacterial burden. These data therefore suggest the possibility that depletion of ILC2s may favor reduced bacterial burdens p.i. To directly test this possibility, we used iCOS-T mice, which allow selective depletion of ILC2s after injection of DT while retaining iCOS 1 CD4 1 T cells due to Cd4 Cre -mediated excision of the DTR locus (17). iCOS-T mice or littermate controls without Cre or DTR expression (referred to as WT) were injected i.p with DT daily from 4 d prior to infection and during infection (Fig. 3C). This led to a reduction in the percentage and absolute numbers of ILC2s in the lung (Fig. 3D, 3E), with no reductions in CD4 1 T cell or ILC1 percentage or absolute numbers (Supplemental Fig. 3CF). Reduction of ILC2s following DT treatment led to a significant reduction in bacterial burden seen in the lung of iCOS-T mice, although we did not observe altered systemic bacterial burdens in spleen and liver using this inducible model (Fig. 3F). Thus, these data suggest that reducing ILC2 numbers may be a strategy for reducing bacterial burden locally in the lung after respiratory F. tularensis infection.

IFN-c in the infected lung environment is important in driving infection-induced reduction in ILC2 numbers
To dissect the mechanistic basis of the observed decrease in ILC2s during F. tularensis LVS infection we next considered whether a soluble mediator present in the F. tularensis LVS-infected lung environment could be important in driving the reduction in ILC2 numbers.
To this end, we developed an in vitro culture system where sortpurified ILC2s were cultured with either naive or LVS-infected cell-free lung supernatant. When ILC2s were cultured with lung supernatants from F. tularensis LVSinfected mice, the number of live ILC2s was significantly decreased over time compared with ILC2s cultured in lung supernatant from noninfected mice (Fig. 4A). Thus, these results suggest that infection with F. tularensis LVS induces the production of a soluble mediator that can reduce ILC2 numbers.
A hallmark of the immune response against F. tularensis LVS is the production of IFN-g (20). As IFN-g has previously been demonstrated to inhibit ILC2 function and activation (21,22), and its production is enhanced after i.n. F. tularensis LVS infection (23), we next considered whether ILC2 numbers were reduced as a result of IFN-g production during infection with F. tularensis LVS. Indeed, as previously described (23), we observed significantly enhanced production of IFN-g by NK cells by day 5 p.i., both in frequency and cell number (Fig. 4BD), a time point that coincides with the observed reduction in ILC2 numbers (Fig. 1E, 1F). Although the frequency of T cells producing IFN-g was also elevated at day 5 p.i., there was no difference in the total numbers of T cells producing IFN-g p.i. (Fig. 4EG).
To determine whether this F. tularensis LVSinduced IFN-g production was sufficient to drive a reduction in ILC2s, we performed our in vitro culture assay in the presence of an anti-IFN-gblocking Ab. Indeed, we found that blocking IFN-g significantly reversed the effects of lung supernatant from F. tularensis LVSinfected animals in decreasing ILC2 numbers (Fig. 4H). Thus, taken together, these data suggest that the production of IFN-g during infection with F. tularensis LVS contributes to the observed reduction in ILC2 numbers.
Our data suggest that expansion or transfer of ILC2s impedes effective immunity to F. tularensis LVS. Whereas IFN-g may impair ILC2 responses, ILC2 expansion has been reported to suppress NK cell responses and protective type 1 immune responses (24). Thus, we next aimed to determine whether increasing ILC2 numbers prevented the production of early innate sources of protective IFN-g. Thus, we analyzed IFN-g 1 NK cell levels after i.n. transfer of ILC2s and F. tularensis LVS infection. Although the proportion of cells producing IFN-g was not altered by infection (Fig. 4I), there was a significant reduction in the total number of IFN-g 1 NK cells in the lung of infected mice that received transfer of ILC2s (Fig. 4J). Thus, taken together, these results suggest that bidirectional crossregulation between IFN-gproducing NK cells and ILC2s is a critical determinant of the immune control during the early stages of F. tularensis infection.

Blockade of IL-5 reduces bacterial burden after F. tularensis LVS infection
Next, we sought to determine potential mechanisms by which ILC2s may suppress the control of bacterial burden p.i. with F. tularensis LVS. Following F. tularensis LVS infection, we detected only minimal expression of the type 2 cytokine IL-13 by ILCs (Supplemental Fig. 4A, 4B). In contrast, using Red5 reporter mice, which a have targeted knock-in of the fluorescent reporter tdTomato downstream from the mouse Il5 promoter (18), we observed high frequencies of IL-5producing ILC2s even in the lungs of both naive and infected mice (Fig. 5A, 5B). In contrast, we observed minimal expression of IL-5 by T cells in the lung in both naive mice and during infection (Supplemental Fig. 4C, 4D), indicating that the major source of IL-5 is ILC2 derived rather than adaptive cells at this time point. These data suggest that IL-5 production by ILC2s may be a potential mechanism by which these cells limit host clearance of bacteria during F. tularensis infection.
To determine whether early sources of IL-5 perturb host responses to F. tularensis LVS infection, we treated mice before and during infection with an antiIL-5 Ab and measured bacterial burdens. We found that treatment with antiIL-5 Ab caused a significant decrease in both pulmonary and liver bacterial burden compared with mice treated with control IgG (Fig. 5C), although no differences were observed in the spleen. Taken together with the data above, these results suggest that ILC2s may suppress bacterial clearance from the lung during F. tularensis infection at least in part via IL-5 production.

Discussion
Much of the current understanding of the immune response during infection with F. tularensis has focused on the role of myeloid cells such as macrophages and neutrophils (6,9), with the role of innate lymphocytes poorly defined. Despite the well-documented role for ILCs to respond rapidly to a variety of immunological challenges (10), infection with F. tularensis LVS only begins to impact the lung ILC compartment with a subset-specific reduction in ILC2s from day 5 p.i., associated with a concurrent increase in ILC1s. Indeed, these changes to the lung ILC compartment have been observed in the context of other intracellular infections, where plasticity and subsequent emergence of an ILC1-like IFN-gproducing cell type via trans-differentiation of an ILC2 population was shown to be important in viral clearance (13). It is unclear whether the changes in the numbers of these specific ILC subsets represent a similarly plastic behavior by ILC2s during infection with F. tularensis, and whether the increase in ILC1s contributes to protective immunity. Evidence that would support the conversion of ILC2s to an ILC1 population in the context of F. tularensis infection is that IL-12 is readily detectable from 72 h p.i. with F. tularensis (25), and this cytokine is critical for the conversion of ILC2s to an ILC1-like phenotype (13). Conversely, infection with F. tularensis could instead propagate the expansion of a minor bona fide ILC1 population, given that IL-12 is also critical for the activation and expansion of ILC1s (10,26). It is therefore important for future work to determine the exact relationship between ILC1s and ILC2s in the context of infection with F. tularensis LVS. Of note, for consistency between experiments, we performed all infection experiments in female mice. Given that there have been reports of some sex differences in lung ILC2s (2730), it will be interesting in future studies to determine whether the reduction of ILC2s during F. tularensis infection are sex specific.
ILC2s themselves are important in the immune response against several intracellular infections (11), with a loss of ILC2s significantly impacting both the integrity of airway epithelium and lung function during influenza infection (12). It is also interesting to note that p.i., ILC2s can provide a source of amphiregulin (AREG) for wound healing and tissue repair (12). However, given the lethality of infection with F. tularensis LVS in mice (15), the role of ILCs in the resolution phase cannot be examined in this particular model. ILC2s can also play a detrimental role in the control of pulmonary infections, contributing to the exacerbation of tissue pathology (11). Indeed, activation of ILC2s results in a strong type 2 immune response characterized by production of the effector cytokines IL-5 and IL-13 (10). Specifically, ILC2-derived IL-13 drives airway hyperreactivity during multiple viral infections (3133). Interestingly, IL-13 can induce an alternative activation state in macrophages in the context of infection with F. tularensis LVS, resulting in enhanced intracellular replication and survival of bacteria (34). However, our data show that ILCs produce minimal IL-13 during F. tularensis LVS infection, suggesting that other alternative cytokines are responsible for the immunoregulatory role of ILC2s in this context, and that non-ILC2 sources of IL-13 are important in promoting alternative activation of macrophages.
In contrast, we show that ILC2s produce IL-5 both at homeostasis and during F. tularensis LVS infection, and that neutralization of IL-5 leads to enhanced bacterial burdens following infection. These data suggest that IL-5 may drive suppression of protective immunity. It has previously been shown that ILC2-derived IL-5 supports the homeostatic function of eosinophils (18), and activated ILC2s can drive the induction of eosinophilic airway inflammation (35). Interestingly, ILC2-derived IL-5 and activation of eosinophils can inhibit the production of IFN-g and the cytotoxic function of NK cells in antitumor immune responses (24). Thus, it is possible that ILC2-derived IL-5 acts to limit type 1 immunity during F. tularensis LVS infection via effects on eosinophils that then suppress NK cells, although more work is required to investigate this possibility. A potential role for ILC2s and IL-5 during F. tularensis LVS infection is in line with a recent study suggesting their importance in regulation of IgM production by B cells in responses to F. tularensis LVSderived LPS vaccination in murine models (36).
A hallmark of infection with F. tularensis is the production of IFN-g (20), which is known to inhibit ILC2 function and proliferation (21,37,38). More specifically, mice with elevated levels of constitutive IFN-g display diminished ILC2 responses during infection with the nematode Nippostrongylus brasiliensis (21). This IFN-g mediated effect is also observed during coinfection with N. brasiliensis and Listeria monocytogenes, a bacterium that provokes a strong IFN-gmediated response (20). Furthermore, the administration of A C B FIGURE 5. Ab-mediated blockade of IL-5 reduces bacterial burden in the lung and liver after F. tularensis LVS infection. (A and B) Red5 reporter mice, which a have targeted knock-in of the fluorescent reporter tdTomato downstream from the mouse Il5 promoter (18), were infected with 1000 CFU of F. tularensis LVS i.n. or given PBS as a control. On day 5 p.i., cells were isolated from lungs, treated with PMA, ionomycin, and protein transport inhibitors, and expression levels of Il5-tdTomato in ILC2s were determined by flow cytometry. (A) Representative flow cytometry plots. (B) Data pooled from two independent experiments (n 5 8). (C) C57BL/6 mice were injected i.p. with either 100 mg of control IgG or anti-IL-5 Ab 4 and 2 d before infection, infected with 1000 CFU of F. tularensis LVS i.n., and injected i.p. with control or antiIL-5 Ab on the same day as infection, injected with Ab again at 2 d p.i., and then bacterial burdens in the lung, liver, and spleen of mice were analyzed at 4 d p.i. Data are pooled from two independent experiments (n 5 8). Statistical analysis was performed using a MannWhitney U test (B and C). *p < 0.05, **p < 0.01.
(I) and absolute cell numbers (J) of IFN-g 1 NK cells at day 7 p.i. in mice receiving PBS or purified ILC2s (as in Fig. 3B). Data are pooled from two independent experiments (n 5 8). Statistical analysis was performed using (A) paired t tests to compare data at each individual time point, (C and D) a KruskalWallis test with Dunn's correction for multiple comparisons, and (F and G) one-way ANOVA with HolmSidak's correction for multiple comparisons. Statistical analysis in (H) was determined at each individual time point by repeated-measures one-way ANOVA with HolmSidak's multiple comparison tests. Significance is shown as LVS-infected versus naive (below red line), and LVS-infected 1 anti-IFN-g versus LVS-infected (below blue line); (I) MannWhitney U test; (J) unpaired t test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. IFN-g during rhinovirus infection significantly attenuates immunopathology, diminishing the output of IL-13 from ILC2s (39). We now reveal that IFN-g also drives an F. tularensis infectioninduced reduction in ILC2s. Enhanced levels of IFN-g observed at day 5 p.i. in our model coincide with the reduction in ILC2 numbers in vivo, which, coupled with our in vitro data showing that IFN-g can reduce ILC2 numbers, highlights the importance of this cytokine in control of ILC2s during F. tularensis LVS infection. Nevertheless, other cytokines such as the type 1 IFNs and IL-27 can also inhibit ILC2 function and survival (20,30). Thus, future work is required to determine whether other soluble mediators can also impact ILC2 numbers during F. tularensis infection.
Taken together, our findings suggest that ILC2s in the lung may impair optimal immunity to F. tularensis LVS. Numbers of ILC2s are reduced during infection, which may be a result of cross-regulation with the protective type 1 immune response that acts to reduce the potentially detrimental role of ILC2s. Mechanistically, our data suggest a potential cross-regulation between NK cells and IFN-g and ILC2s and IL-5 in determining the magnitude of F. tularensis LVS bacterial burdens during the early stages of infection, with perturbation of this axis via cytokine neutralization or manipulation of ILC2 abundance having dramatically altered the pathogen burden.
Overall, our data begin to highlight the potentially detrimental role of ILC2s in the control of infection with F. tularensis. Future work focusing on how this subset-specific reduction in ILC2s occurs may allow for a more targeted approach of manipulation of ILC2 numbers. Although more work is required to determine the overall outcome of depleting ILC2s on the broader immune response and the host, our work highlights that early depletion of ILC2s is an important pathway to explore further with the aim of promoting a more beneficial host immune response against F. tularensis.